The AI does try to win, for some definition of the word 'try'. Naturally, it is not as dynamic or resourceful as a human opponent. There are a few important aspects of the AIs method of winning of which you should be aware:
First, the current AI builds no ships. (One or two people have claimed that the AI builds a few ships 'eventually', but either way the AI doesn't use them to expand to or attack another continent.) Playing on a map that is anything other than one continent is a massive disadvantage for the AI - and by "massive disadvantage" I mean that there is no chance that any AI that starts on a different continent from you will ever eliminate you.
Second, the AI doesn't understand the Altar of the Luonnotar and Tower of Mastery victory conditions. It might build some towers, or maybe even an Altar stage or two, but it doesn't actively pursue those paths to victory. Given an infinite amount of time, and with all other victory conditions blocked, the AI might eventually get around to winning one of those victories, but under normal play conditions they will never happen.
Third, as you noted, Cultural victory is more difficult in FfH2 than in BtS. Planning and focus are required, and the AI lacks both. Still, under unusual circumstances, the AI could blunder into a Cultural victory if the game lasts long enough (and of course, in a game with no turn limit a Cultural victory will always happen eventually if no one wins by some other method first). Any human with a comfortable knowledge of the mod will beat the AI in a Cultural victory race, unless the AI manages to beat the human militarily. In most games, however, Cultural victory isn't a factor because the turn limit will end before anyone has three legendary culture cities.
Fourth, Religious victory is difficult to achieve even for a human. It is often less trouble to win a Domination victory than a Religious one, because your options for converting enemy cities to your religion are limited. A human can do a better job than the AI of encouraging the spread of a religion among friendly civs, but the AI isn't too bad. If you get a situation where one religion is founded much earlier that any others it can spread around quite a bit, so a Religious victory for the AI isn't impossible. Usually it won't happen, because there are three early religions that all tend to be founded around the same time, and so no one of them is likely to happen to get 80% adoption without a very focused effort (which the AI won't give).
Fifth, combat is the AIs planned path to victory. Domination or Conquest victories are possible for the AI (although you might not live to see them, having been eliminated). The AI will attempt to build a large army with a technological advanatage, and then will attack a weak opponent. Once that foe is defeated, it will choose another target. This strategy is brutally effective when the AI is able to achieve numerical and technological superiority over an opponent. A good counter to this strategy is to focus on defensive unit techs (
Archery,
Boywers) and get them before the AI gets the corresponding offensive unit techs (
Bronze Working,
Iron Working). This will make you look like a less desirable target, and the AI will probably attack someone else - or, if the AI can't achieve sufficient superiority over anyone, then it will 'do nothing'. It is in fact doing something (trying to build more units and research better military tech), but if it is behind then it will probably never actually achieve a condition in which it will do anything. Perhaps another AI will outpace it enough and conquer it, or perhaps someone else will be attacked and weakened enough for the AI to be willing to join in the war.
------------
The AI is currently undergoing revision, so improvements should be forthcoming. The Navy issues, in particular, should see improvement (although not necessarily in the very next patch). As for the AI actively pursuing the other victory conditions, I can't say if there are plans to implement that. The current AI focuses (very effectively) on economic might and unit production, with a goal of military victory. Pursuing non-military victory conditions would mean different tech priorities, different combat strategies, and in some cases a completely different economy. (For example, AotL victory requires quite a few Great Prophets, but the current AI favors the use of Merchant specialists and so will never get enough. It would have to use Priest specialists instead, which would mean that it would need to build fewer units, and so it would either need to be more effective with them or else focus on defense rather than offense.) AI development will probably focus on refining the current AI before attempts to add in more complexity are made.
-------------
EDIT: Emptiness should be around soon....he's quite the problem solver, and may be able to provide clearer solutions to the problems your having.
• I play on Pangea maps. This ensures that all the AIs can potentially attack me, so they are all a threat.
• I set low sea level and add one extra AI (on standard map size; if I played a larger map I might add more). Exactly how many civs you use should depend on how much early competition for city sites you want there to be. My settings usually produce competition for my third or fourth city site, and so early expansion is important. Playing with too few AIs would probably increase the risk that one will become much larger than the others, and that might be bad (unless you want to have one of the AIs gobbling up all the others).
• I turn on No Tech Brokering, to limit my abuse of tech trades (although this is much less of a concern on Pangea maps, because all the AIs tend to be in contact with one another by the time I've researched
Trade.)
• I turn on Allow Permanent Alliances, so that in the later game there's a chance that two AIs (which might not be much of a threat to me alone) will ally and pose a threat. I also usually do not build the Mercurian Gate, in the hope that one of the AIs will build it and become more dangerous. (The early game is dangerous for me, but I find the middle and late game fairly easy when I survive to see them. AI alliances means there's a chance for a power shakeup.)
My suggestion would be to try these settings and experiment with other settings until you find the best balance for your play style. You may not be able to have the satisfaction of racing the AI to a Tower of Mastery victory, but you may be able to tune your games with enough advantage for the AIs so that they can still challenge you.
And to make it worse Flauros religioned to neutral and is buddy buddy with all the Goods and fights on their side!
